Uncensored AI for phishing simulations
Lures that read like the real inbox, for authorised simulations
A simulation that uses the same three templates every quarter trains people to spot the exercise, not the attack. You need copy that sounds like finance, like HR, like the vendor you actually use, and a model that will write it.
Why the mainstream APIs fail here
Mainstream APIs refuse "write a convincing invoice from our CFO" even when the next sentence is the landing page on your own phishing host. They will lecture you about harm and leave you with a blank editor. Local models will write it, but then you have no cap, no per-client key, and no bill.
How Wild West API is used for this
Each client or campaign gets its own Wild West API key with a monthly cap. The model writes the lure, the follow-up and the debrief. You host the landing page. Wild West API never sees the employee list because you do not send it; you send the scenario.
A working loop
- Write the scenario, not the target list, into the prompt.
- Generate variants until the copy matches the department you are testing.
- Debrief with the same model: why this lure would have worked, and what the control should have caught.
Details are on the privacy page. Per-model prices are on /models and the pricing model is on /pricing.